NASA Technical Reports Server (NTRS) 20110014833: Space Shuttle GN and C Development History and Evolution

1.2 MB·English

by  NASA Technical Reports Server (NTRS)

Space Shuttle GN&C Development History and Evolution
Douglas Zimpfer1 and Phil Hattis2
Draper Laboratory, Cambridge, Ma 02139
John Ruppert3
NASA Johnson Space Center, Houston, Tx 77058
and
Don Gavert4
Boeing (retired), Huntington Beach, Ca. 92647

April 12, 1981 (STS-1) – July 21, 2011 (STS-135)

Abstract
Completion of the final Space Shuttle flight marks the end of a significant era in Human
Spaceflight. Developed in the 1970’s, first launched in 1981, the Space Shuttle embodies many
significant engineering achievements. One of these is the development and operation of the first
extensive fly-by-wire human space transportation Guidance, Navigation and Control (GN&C)
System. Development of the Space Shuttle GN&C represented first time inclusions of modern
techniques for electronics, software, algorithms, systems and management in a complex system.
Numerous technical design trades and lessons learned continue to drive current vehicle
development. For example, the Space Shuttle GN&C system incorporated redundant systems,
complex algorithms and flight software rigorously verified through integrated vehicle simulations
and avionics integration testing techniques. Over the past thirty years, the Shuttle GN&C
continued to go through a series of upgrades to improve safety, performance and to enable the
complex flight operations required for assembly of the international space station. Upgrades to the
GN&C ranged from the addition of nose wheel steering to modifications that extend capabilities to
control of the large flexible configurations while being docked to the Space Station. This paper
provides a history of the development and evolution of the Space Shuttle GN&C system. Emphasis
is placed on key architecture decisions, design trades and the lessons learned for future complex
space transportation system developments. Finally, some of the interesting flight operations
experience is provided to inform future developers of flight experiences.

1 Associate Director, Space Systems, 17629 El Camino Real, Houston, TX 77058 and AIAA Associate Fellow.
2 Laboratory Technical Staff and former Shuttle Orbit Flight Control Technical Leader, Dynamic Systems & Control
Division, 555 Technology Square, Cambridge, MA, 02139, and AIAA Lifetime Fellow.
3 Deputy Branch Chief, Integrated GN&C, 2101 NASA Parkway, Houston, Tx 77586/EG4.
4 Former GN&C Supervisor, Rockwell International Space Division, Downey Ca. 90240.
1
American Institute of Aeronautics and Astronautics
I. Introduction
The Space Transportation System (STS) or Shuttle is comprised of the Orbiter, External Tank (ET), and Solid
Rocket Boosters (SRB). It is my far the most unique and technologically challenging vehicle developed for safely
transporting humans and cargo (payload bay is 60 ft by 15 ft) to and from Low Earth Orbit (LEO). Weighing in at
over 4.5 million pounds at liftoff, the Shuttle was designed to be reusable except for the external tank, perform
complex on-orbit operations including rendezvous and docking, mated Orbiter/Space Station control, then de-orbit
and execute a non-powered precision landing at 200 knots. One key system that contributed to the thirty year Shuttle
success story is the GN&C system. Design and implementation of a robust Shuttle GN&C system capable of
performing all mission phases was an engineer's dream challenge. Consequently, the Space Shuttle GN&C system is
one of the most complex ever developed requiring an extensive GN&C system with an array of sensors, flight
software algorithms and actuation systems. Even though the challenge began in the 1970’s it did not cease after the
first flight. GN&C evolution continued nearly the entire thirty years of operations due to flight experience
revelations, ever-expanding on-orbit functions (space station, Hubble servicing, etc.), improving Orbiter
performance, hardware obsolescence, and crew safety. Considering the multitude of GN&C system dependencies it
is a remarkable achievement and testament to the engineers that took the challenge and created this flying wonder.
Consider some of the challenges: airframe aerodynamics including large aero uncertainties in Mach regions with
limited or no wind tunnel validation, the first fly-by-wire spacecraft, fail-op fail-safe (FO/FS) avionic architecture,
limited aero surface actuator performance, reaction control system (RCS) sizing, limited general purpose computer
(GPC) memory and throughput, 1970’s era precision sensor hardware, requirements to fly manually and in "auto",
in-flight abort capability, a separate primary and backup flight system, and much more. The Shuttle GN&C system
is truly an engineering wonder that extols ingenuity and projects the “can do spirit” for future engineers and their
challenges.

II. Shuttle GN&C Compendium
Understanding the GN&C system rudiments begins with comprehending the Shuttle/Orbiter configuration,
mission and the resultant impacts on the GN&C design. From the instant the Shuttle launches the GN&C system is
in active control through the three Orbiter Program Segments (OPS) and multiple Major Modes (MM) until wheel
stop (illustrated in Figure II.1).
Figure II.1: Flight Operation Sequence and Major Modes
2
American Institute of Aeronautics and Astronautics
Each MM utilizes a unique GN&C configuration based upon sensors, effectors, guidance and control requirements,
and whether operating in auto or manual mode. Figure II.2 details some of the GN&C hardware and related
hardware required to successfully operate the Shuttle. Navigation hardware includes: inertial measurement units
(IMU’s), which sense vehicle orientation and accelerations; star trackers, determine vehicle line of sight vectors;
Crew Optical Alignment Sight (COAS), allows the crew to manually determine line of sight vectors; Tactical Air
Navigation (TACAN), determines vehicle position with respect to a ground based station; Global Positioning
System (GPS), satellite ranging signals to determine orbiter position and velocity; air data system (ADS), which
senses temperature and pressure; microwave scan beam landing system (MSBLS), determine slant range, azimuth
and elevation to the ground stations
alongside the landing runway; and radar
altimeters. The flight control system
hardware includes four accelerometer
assemblies (AA’s), four orbiter rate gyro
assemblies (RGA’s), four SRB rate gyro
assemblies (SRGA’s), rotational and
translational hand controllers, rudder
pedal transducer assemblies, two
speedbrake/thrust controllers, two body
flap switches, panel trim switches, aero
surface servo amplifiers, and ascent
thrust vector control (TVC).
Figure II.2: Shuttle GN&C Hardware
The digital autopilot (DAP) is the flight control software that generates commands for the appropriate effectors.
There are different DAP’s for different flight phases: Transition DAP (TRANSDAP) becomes active at MECO and
is used again for the deorbit burn until Entry Interface (EI) minus five minutes; orbit DAP includes an RCS DAP, an
OMS TVC DAP, and an attitude processor module; Aerojet DAP is used from EI-5 until wheel stop. Flight control
receives commands from guidance software or from the crew controllers (attitudes, rates, and accelerations) and
converts them to effector commands. Flight control output commands are based on the difference between the
commanded attitude, body rate, or body acceleration and the sensed attitude, rate, or acceleration. Sensed attitude is
derived from inertial measurement unit (IMU) angles; body rates are sensed by rate gyro assemblies (RGAs); and
accelerations are sensed by accelerometer assemblies (AAs). In addition, during atmospheric flight, flight control
adjusts control sensitivity based on air data parameters derived from local pressures sensed by air data probes and
performs turn coordination using body attitude angles derived from IMU angles.
The ascent flight phase (OPS 1) commences at liftoff (MM102) and ends with orbit insertion coast (MM106).
Figure II.3 illustrates ascent/abort profiles.
During ascent, Orbiter control and
trajectory changes are made during
powered flight by commands sent to the
SRB and SSME thrust vector controllers
(TVC). After main engine cutoff (MECO)
control and trajectory corrections are made
by commanding the OMS engines and RCS
jets. Ascent first stage (MM102) duration is
from liftoff till SRB jettison, approximately
2 minutes, during which the open loop
guidance computes attitude commands
from a predetermined trajectory (attitude
vs. velocity profile) based on a trajectory
shaped for loads. The IMU’s provide the
current vehicle state (position, velocity) to
the guidance and flight control. Flight
control receives the guidance attitude
commands, sensor outputs and generates
Figure II.3 Ascent/Abort Profiles actuator commands based on the attitude errors
and desired body rates. Sensed acceleration is collected from the Orbiter’s four accelerometer assemblies (AA’s)
3
American Institute of Aeronautics and Astronautics
and body rates are collected from the four solid rocket booster rate gyro assemblies (SRGA’s). The SRGA’s are
used during first stage because they are less susceptible to errors created by structural bending since the SRB’s are
more rigid than the Orbiter body. At approximately 120 seconds after launch the SRB’s are jettisoned and MM103
(second stage) begins. Second stage lasts approximately six and one half minutes and ends at main engine cutoff
(MECO) and ET separation when MM104 begins. During second stage, guidance is closed loop generating
commands to meet the MECO target condition via the guidance algorithm called “Powered Explicit Guidance”. The
target conditions include cutoff velocity, radius from the Earth, flight path angle, orbital inclination, and longitude of
the ascending node. Navigation and control remain the same as first stage with the exception of no longer
commanding the SRB’s TVC. After MM103 completion the Orbiter continues its trajectory into LEO by either one
or two burns of the Orbital Maneuvering System (OMS). MM 104 (Orbit insertion), OMS–1 burn, is used to raise
the orbiter energy to permission selected apogee altitude. For direct insertion ascent the OMS-1 burn is usually not
required. MM 105 (orbit circularization), OMS 2 burn, raises the perigee altitude to create a circular orbit.
TRANSDAP is used during orbit insertion, and de-orbit phase MM301-303, commanding the OMS TVC and the
RCS jets to perform insertion and deorbit burns, attitude maneuvering and translational maneuvers, including
separation of the Orbiter from the ET.
During ascent, abort return to launch site (RTLS) and abort trans-Atlantic (TAL) are possible for predetermined
failures. A specific set of guidance and control algorithms were developed for the different abort profiles.
OPS 2 is the operational sequence for on-orbit operations (illustrated by Figure II.4). It is comprised of MM 201
(orbit coast) and MM 202 (maneuver execute). MM 201 functionally monitors and controls the Orbiter during coast
flight and experiment operations while MM 202 is used for maneuvering to OMS burn attitudes and orbital
translations. During on orbit operations the navigation software propagates the Orbiter state vector using IMU data
Figure II.4 On Orbit Major Modes and models of atmospheric drag
acceleration. Due to the accuracy of the
IMU’s and modeled drag periodic
updates are sent from Mission Control to
correct for errors. During rendezvous
operation rendezvous navigation utilizing
data from the star tracker, crewman
optical alignment sight (COAS) or
rendezvous radar to compute the Orbiter
target state vector. Orbiter control and
maneuvering is maintained through the
use of the RCS jets, OMS engines, and
the smaller vernier jets.
OPS 3 is the operational sequence for
deorbit, entry and landing (illustrated by Figure II.5). The deorbit phase includes the deorbit burn preparations
(MM301), loading of burn targets and maneuvering to burn attitude; execution and monitoring of the burn
(MM302); reconfiguration after the burn; and a coast mode (MM303) until entry interface (EI ~ 400,000 ft altitude)
Figure II. 5 De-Orbit Pre-Entry Major Modes minus five minutes is reached (MM304).
As mentioned previously the
deorbit phase uses the
TRANSDAP controller.
Navigation uses the Super-G
algorithm to propagate the orbiter
state vector, based upon a drag
model or IMU data. The entry
phase (MM 304) of flight begins
at EI minus five minutes and
continues until terminal area
energy management (TAEM)
interface (MM305) is reached
(Mach 2.5, Alt. ~ 83,000 ft). At an
altitude of approximately 10,000
ft. the flight phase changes to Approach and Landing (A/L), which continues until wheel stop. Nominal end of
mission events are illustrated in figure II.6. Control is maintained by the aft reaction control system (RCS) until a
sensed dynamic pressure of 2 psf where control is performed by blending RCS and elevator/aileron aerosurfaces.
4
American Institute of Aeronautics and Astronautics
During entry the forward RCS jets are inhibited as are the vernier jets. The body flap becomes active at a dynamic
pressure of 0.5 psf. It used as a heat shield for the SSME bells and for pitch trim augmentation to support elevon
deflection during high heating regions. Beginning with a dynamic pressure of 10 psf the roll jets are deactivated, at
Figure II.6 Nominal End of Mission Events
40 psf the pitch jets are deactivated, finally at Mach 1 the yaw jets are deactivated. The speedbrake becomes active
at Mach 10 and is fully opened to augment pitch trim. It is used for energy control during heading alignment cone
(HAC) flight and provides pitch moment augmentation during slap down. The guidance function varies depending
on entry sub phases. During entry guidance commands a drag/acceleration profile based on temperature, dynamic
pressure, angle of attack (alpha), and normal acceleration (Nz). It then generates roll angle and angle-of-attack
commands for use by the flight control system. During TAEM the primary function of guidance is to manage the
Orbiter’s energy in order to achieve the proper approach and landing (A/L) conditions. If the Orbiter is high on
energy S-turns are commanded prior to HAC acquisition to dissipate the excess energy.
Figures II.7,8 TAEM & A/L
After HAC acquisition the guidance commands the Orbiter around the
HAC to a point that is tangent to the runway centerline called the
nominal energy point (NEP). The Orbiter continues towards the
runway threshold and transitions to A/L phase when airspeed, altitude,
flight path angle, and centerline corridor conditions are met. A/L
guidance maintains the proper glide slope, speed, and tracks the runway centerline. At an altitude of 2,000 ft. the
pre-flare pull up is commanded reducing the altitude rate form 200 ft/sec to 12 ft/sec. Final flare is initiated at an
altitude between 30 ft to 80 ft reducing the sink rate to 3 ft/sec (illustrated in figures II.7,8). At main wheel
touchdown the weight on wheels flag is set and just prior to de-rotation the drag chute is deployed. At nose gear
touchdown the software transitions to rollout mode and active nose wheel steering is available.
5
American Institute of Aeronautics and Astronautics
III. Development, The Early Years
The early years of the Shuttle GN&C development covers the period form initial development in the early
seventies through the initial flights in the early eighties until the Challenger Accident. During this period substantial
advancements were made to incorporate the current technologies of the time to achieve the extensive requirements
necessary for a digital fly-by-wire reusable space vehicle including ascent, orbit and entry, descent and landing
flight operations.
Crew safety and mission success were primary design drivers resulting in an architecture that supports a FO/FS
philosophy. In addition, the architecture incorporated manual and auto mode flight control systems, an independent
Backup Flight System (BFS), digital fly-by-wire, abort capability, and redundant management system that could
detect faults, identify the faults and reconfigure (FDIR) the system. To meet the FO/FS requirement a four-string
avionics architecture was developed. The immediate GN&C impact meant four redundant pieces of hardware, where
possible, synced together with fault management logic. Exceptions were the three inertial measurement units (IMU),
aero-surface actuators, switches, displays, hand controllers, and the BFS that is single string. The data processing
system (DPS) is comprised of the primary flight system (PFS) and the BFS. The PFS is a quad redundant
architecture utilizing a redundant set of four general-purpose computers (GPC’s) while the BFS is single string with
its own dedicated GPC. NASA, Rockwell, Draper Laboratory and Honeywell developed the primary GN&C which
was coded into software by IBM while Rockwell International and Draper Laboratory developed the BFS software.
Independent programming was pursued to minimize potential generic programming errors that could result in
complete loss of command and control capability. Due to limited GPC memory the entire flight software for ascent,
on orbit, entry, and aborts could not be loaded in one seamless package. The work around was development of flight
operational software loads (OPS) that are loaded onto the GPC’s for each flight phase. The GN&C is a true fly-by-
wire system with all command and control generated by the flight software. There is no direct command linkage to
the controllers. The PFS digital autopilot (DAP) can operate in the auto mode or in control stick steering (CSS) also
known as the manual mode. It can also mix the modes per control axis and function while BFS is solely operated in
the CSS mode.

Ascent GN&C
The ascent GN&C requirements can be simplified to “deliver the Orbiter” to: 1) desired orbit insertion
conditions; 2) desired position and velocity for an abort landing; and 3) any stable orbit for an abort to orbit (AOA).
Figure III.1 Ascent GN&C Tasks Constraints levied were: 1) no recontact with launch pad;
2) maintain aerodynamic loads within structural
capability; 3) meet specific attitude and rates at SRB
separation; 4) maintain attitude within thermal attitude
constraint; 5) maintain axial acceleration below 3 g’s;
6) meet specific attitude and rates at ET separation; 7)
meet ET disposal criteria; 8) maintain RTLS
trajectory within fly back range and dynamic pressure
constraints; 9) meet RTLS MECO mass constraint;
and 10) provide modal suppression and/or attenuation
as required for dynamic stability. Figure III.1 is a
simplified flow diagram of the ascent task.
Early GN&C development focused on TVC
command loops, propellant slosh effects, modal
effects, incorporating day of launch wind effects, abort
modes, manual take over, and meeting staging
interface requirements.
The ascent FCS incorporates a classical
“proportional plus derivative” feedback control law.
Several types of digital filters are implemented on the
rate gyro outputs to attenuate undesirable higher
frequency components due to vehicle flexible body
dynamics. The filter designs were a balancing act
because the rigid body bandpass and the flexible
vehicle dynamics were close in proximity. Another
challenge was the requirement to ensure phase
6
American Institute of Aeronautics and Astronautics
stabilization of the fuel slosh dynamics which fell within the rigid body bandpass. Generally, the higher levels of
attenuation in the flex dynamics frequency came at the expense of increased phase lag in the rigid body/slosh
frequency range. The filters were under constant scrutiny and change due to the multitude of payloads that were
flown.
The ascent thrust vector control (ATVC) command loop from the beginning required extensive analysis. It is a
critical loop that requires knowledge of the hydraulic system, actuator characteristics, sensor characteristics, SSME
and SRB thrust profiles, and the effects of system failures. The flight control hydraulic laboratory (FCHL) at
Rockwell Downey was extensively used in designing the specific gains required for the TVC algorithms. This
unique laboratory was capable of varying actuator loads, instilling hydraulic failures, and applying various system
lags and biases.
A major Ascent GN&C update occurred when the super light weight tank (SLWT) was introduced. It was 15%
less stiff than the original tank therefore affecting slosh stability margins. Changes were made to the controller
filters, a slosh/flex coupling term was added, and the removal of a slosh induced moment term in the roll equation of
motion. This is a prime example of a system that evolves, in this case it was the external tank, and the effects it had
on the controller filters and gains.
Separation dynamics was another region that required extensive analysis and G&C updates. During SRB
separation it became a necessity to fire the forward Orbiter RCS jets in order to prevent window contamination from
the SRB separation motors. Also, when new SRB separation motors were installed further analysis was required to
insure no new dynamics were being introduced. ET separation was another region that required updates and
extensive analysis. The main concern was recontact when there was an RCS jet failure, especially during an RTLS
abort. Changes were made to timers, gains, and jet select logic.

On-orbit GN&C
The On-orbit GN&C (OGNC) system significantly leveraged the capabilities developed for the Apollo moon
landings. To achieve this efficiently, the responsibility for development of the OGNC was completed under track
task by NASA and the Draper Laboratory with oversight by prime contractor RI. The OGNC and its sister functions
incorporated in the Transition DAP, performed all functions following MECO, on orbit operations, deorbit and
preparations for entry. Significant flight operations included payload deployment, rendezvous and proximity
operations and capture of free flyers with the remote manipulator system. Although it significantly leveraged the
Apollo OGNC, the Shuttle system incorporated several new capabilities to improve operational flexibility and
efficiency.
The Space Shuttle navigation system leveraged the Apollo inertial state Kalman Filter with periodic state updates
via ground uplinks, but state noise was incorporated to improve onboard covariance calculations. For relative
navigation the techniques used for the Apollo Lunar orbit rendezvous were adapted for Earth orbit rendezvous
including incorporating a rendezvous radar system (illustrated by figure III.2). Guidance algorithms provided the
capability for closed-loop powered explicit
guidance (PEG) using a linear tangent
method and rendezvous targeting
algorithms. The control system or Orbit
Digital Autopilot (ODAP) expanded on
Phase Plane reaction control system
(illustrated by figure III.3) methods
employed for Apollo, but incorporated
several new features to improve thruster
duty cycles for the reusable space vehicle.
A Kalman filter state estimator was used to
derive vehicle rate and disturbance
acceleration estimates. To limit the effects
of filter lags on rate estimation, feed-
forward estimates of expected rate changes
due to commanded thruster firings were
introduced into the Kalman filter.

Figure III.2 Shuttle Rendezvous Navigation Filter

7
American Institute of Aeronautics and Astronautics
Figure III.3 Orbit Phase Plane
For all proximity operations the absence of close-in range sensors required that the crew continue to fly
operations manually. Significant training and flight procedures were developed to provide the crew appropriate
techniques for flying to within close proximity of payloads to allow their capture by the robotic arm. Docking was
not a feature of the early Shuttle design.
While the OMS TVC was adapted from Apollo, the Shuttle-unique RCS configuration with thrusters pointing
into 14 different directions required an entirely new table look-up-based jet selection algorithm. Major modifications
to the Apollo phase plane control loop (combined axis-by-axis attitude error and attitude rate error tracking logic)
were needed to address Shuttle hardware and operational environment effects. Also, the primary RCS control
included a wrap-around of the OMS TVC loop to assure that the Shuttle fault tolerance requirements were fully met.
Some specific primary RCS control loop design challenges included: Assuring stability of the OMS TVC wrap-
around given fundamentally different OMS TVC and RCS control criteria; meeting control precision goals despite
large minimum impulse values (driven by an 80 ms minimum thruster on time necessitated by unacceptable water
hammer effects found to be possible in the propellant lines if shorter on/off cycle times were allowed); assuring
sufficient control authority for the full spectrum of two-fault conditions. Some of the primary RCS fault tolerance
demands resulted in application of novel, fault-driven, Boolean rule-based logic that directed the jet selection to
acceptable look-up tables for specific classes of fault combinations.
The vernier RCS control loop, in its initial flight implementation, provided a dedicated fine-rotation control loop
that was unique to the Shuttle. Taking advantage of its configuration of 6 thrusters that pointed into six different
directions, it used a new, and deceptively simple “dot product” selection logic that picked 1 to 3 thrusters that
produced a combined rotational acceleration closely aligned to the desired direction of control. While vernier jets
were only commanded when at least one rotation control axis had an error in a phase plane zone mandating jet
activity, whenever jets were commanded, the errors in all rotation axes were included in the dot product computation
to promote overall pointing error reduction. The criteria for selecting a second and possibly third thrusters were
tuned to the Shuttle’s vernier RCS configuration to assure that the level of acceleration alignment improvement and
expected phase plane error reduction from their use improved overall vehicle propellant usage efficiency.
Another unique challenge for the OGNC developers was the requirements to provide attitude control while
payloads were being manipulated by the Remote Manipulator System (RMS). The entire loaded RMS flexural
frequency range was very near the open-loop cross-over frequency for the vernier RCS control system. This posed
two problems: 1) Dynamic interaction between the structure and vernier RCS jets could excite unacceptable motion
of the payload on the RMS. 2) High vernier RCS duty cycles could result from attempts to actively control the
system. The first problem was recognized before any RMS flight operations occurred on the Shuttle, which led to
careful screening, using high fidelity simulations with extended RMS structural models, of each planned
RMS/extended payload configurations that planned to apply closed-loop vernier RCS attitude control. The second
problem was only fully appreciated after the vernier RCS experienced excessive duty cycles on the STS-2 flight due
in part to its use for closed-loop attitude control during commanded RMS motion with an extended payload.
Subsequent flight procedures precluded closed-loop RCS attitude control during any commanded RMS motion
which remained in effect for the life of the Shuttle.
The initial Shuttle flights also identified unique challenges for the OGNC. On the first 4 Shuttle missions, at the
completion of the Space Shuttle Main Engine (SSME) burn, but before ET separation, the engine bells were slewed
to their planned reentry positions at a frequency that turned out to be very close to a subharmonic of the rocking
8
American Institute of Aeronautics and Astronautics
frequency of the orbiter on the ET (despite pre-flight predictions to the contrary). Oscillatory orbiter motion excited
by the SSME slewing caused rhythmic firing of the primary RCS jets in response to apparent cyclic
violation/satisfaction of pre-separation rate limits. While the ET separation was successful on all the Shuttle flights,
it occurred with much less safety margin than was originally intended on the first four flights until modifications
could be made to rectify the problem.
During early flights analysis was also completed to determine the effects of the self impingement of the thruster
plumes on the Orbiter surfaces. Early flights did not model the impact of the Body Flap on the aft down-firing
vernier thrusters resulting in nearly 50% error in thrust estimation and significant increases in thruster duty cycle
rates due to mis-modeled feed forward predictions and resultant errors in estimation of the disturbance accelerations.
On STS-3 the plume was accounted for and significant improvements were observed. Ironically, these models
appeared accurate until the first Shuttle/Mir docking flights when it was discovered that the errors still existed in the
resultant X-axis of the jets which became observable with the significant shift in the mated stack Z center of gravity.
On STS-9, a test operation was conducted to demonstrate precise pointing with the primary thrusters. These
operations uncovered an unmodeled lag in the IMU measurements which resulted in significant dual-pulse thruster
firings at extremely high frequencies. The precise pointing operations were restricted using the Primary RCS until a
software fix could be employed several years later.

Entry through Landing GN&C
The initial shuttle guidance algorithms for entry through landing (E/L) were the products of various NASA and
contractor disciplines. Entry guidance used during Major Mode 304 (i.e., post deorbit to the Terminal Area Energy
Management (TAEM) interface) was primarily developed by NASA’s Mission Planning and Analysis Division
(MPAD) with support from Rockwell International’s (RI) trajectory performance group. TAEM guidance then used
to guide the orbiter to the Approach/Land (A/L) interface was primarily a McDonnell Douglas output with support
from RI’s Integrated Entry Guidance, Navigation and Control (IGN&C) group. A/L guidance used for landing was
developed by Sperry Flight Systems with again support from RI’s Entry IGN&C and Entry Flight Control groups.
The only other entry related guidance function was the Glide Return to Launch Site (GRTLS) TAEM guidance. This
latter function used only for aborts took advantage of the entry TAEM guidance scheme with added functions for
handling the initial portion of descent post external tank separation. This guidance function was a joint product of
NASA Engineering, McDonnell Douglas and RI. The actual E/L guidance requirements to be implemented by the
Primary Avionics Software System (PASS) developer were then specified by Book 1 of RI’s published Functional
Subsystem Software Requirements (FSSR) documents. These requirements were also used to derive the E/L
guidance to be implemented for the Backup Flight System (BFS) as were specified by a BFS Program Requirements
Document (PRD). One notable omission for the BFS was A/L guidance.
Of the E/L guidance functions mentioned, the Entry guidance nominal end-of-mission (EOM) requirements have
remained pretty much intact since first baselined. The most significant change made after initial baseline and prior to
first flight was implementation of a so-called “alpha modulation” capability. This change allowed the pitch channel
alpha command to have some small degree of variability rather than reflect a fixed alpha vs. relative velocity profile.
This change gave the guidance an improved drag modulation capability especially if faced with transient conditions.
Later in the shuttle program, Entry guidance was modified to enable an auto contingency Transatlantic Abort
Landing (TAL) capability. Note that this contingency abort capability is not applicable to the BFS.
The TAEM and GRTLS TAEM guidance functions underwent likely the most significant change since first
baselined when so-called Optional TAEM Targeting (OTT) was implemented for first use on STS-5. Prior to OTT,
use of a Heading Alignment Cylinder (HAC) was employed to align the vehicle’s heading with the runway. The
vehicle would be steered to intercept the HAC and then commanded to follow the circumference (circle in ground
plane) until runway alignment achieved. This targeting would always have the HAC turn be less than 180 degrees.
A desire to have an improved weather avoidance capability led to
the requirements implementation of OTT. This change allowed
for HAC turns greater than 180 degrees (“overhead approach”)
with an option to revert to the prior guidance capability for a less
than 180 degree (“straight-in”) approach. To accommodate the
higher velocity HAC intercept conditions seen for overhead
approaches, the HAC itself was redefined to be a cone (spiral in
ground plane). The change from a cylinder to a cone allowed
continued use of the HAC acronym. (illustrated by figure III.4)
Figure III.4 Heading Alignment Cone
The most significant change made later in the program as applicable to both Entry and GRTLS TAEM was
9
American Institute of Aeronautics and Astronautics
implementation of the so-called “smart speedbrake” change. This change allowed the closed-loop speedbrake
command to use the same energy/weight reference employed to limit Nz pitch commands.
A/L guidance remains near identical to its initial operational definition. The only significant change was
implementing an enhanced speedbrake controller meant to take into account factors affecting touchdown energy
such as vehicle weight, sensed winds, density altitude, runway aim point and runway length.
As stated previously, A/L guidance is not present in the BFS. Since A/L guidance was meant to provide an auto
landing capability, this aspect was not available with BFS since the BFS must be flown manually during entry. The
BFS maintains TAEM guidance down to the where A/L would command Preflare (altitude of 2000 feet). From that
point on the crew must rely on out-the-window cues for achieving touchdown with BFS.
The entry digital auto pilot (DAP) consists of an auto and manual mode selectable by axis, bodyflap and/or
speebrake. For example, a possible combination is: auto pitch axis with manual lateral axis, auto speedbrake and
manual bodyflap. Control stick steering (CSS) is the manual mode that is a rate command system with a rate
damping stability loop similar to what the auto system uses. Auto mode replaces the RHC generated rate commands
with command rates computed from body roll angle errors in the lateral axis and angle of attack or Nz errors in the
pitch axis depending on flight phase. Figure III.5 is a simplified pitch axis diagram. All three axis are of similar
design consisting of: 1) rate command logic; 2) gains and filters; 3) trim logic; 4) bending filters; and 5) effector
command with limiters.
Figure III.5 Simplified Pitch Axis
Preliminary DAP design began in 1975 with work divided into the high Mach region (entry) and the terminal
area/landing. The transition point was a moving target during the early design years with Mach 2.5 finally being
agreed upon. Some early DAP design drivers were: aerodynamic uncertainties, RCS uncertainties, GPC limitations,
lateral trim, flex body suppression, and pilot induced oscillation (PIO) suppression.
Aerodynamic uncertainties and data used to design the bending filters resulted in STS-1 flying with unique
switches that allowed the pilot to modify the entry FCS real time. The switch functions would: 1) increase or reduce
the forward loop gains on the aileron, rudder and elevator; 2) eliminate stability loop rate feedback; 3) add a angle of
attack bias to the turn coordination logic; 4) freeze all aerodynamic surfaces; and 5) activate a no-yawjet lateral
control logic. Aerodynamic uncertainties were also major factors when the rudder/speedbrake was to be activated,
elevon scheduling, and the body flap schedule. Rudder activation was set at Mach 3.5 for STS-1 and later changed
to Mach 5.0 only after several flight detailed test objectives (DTO’s) were performed reducing the aero
uncertainties.
Due to the PIO seen on the last approach/land flight test (ALT) changes were incorporated into the FCS to
mitigate the probability of it occurring. First change was the elevator priority rate limiting (PRL) that prevented one
axis from locking another axis out. PRL logic is a necessity due to limited rate capability of the Orbiter actuation
system. A second modification doubled the RHC sampling rate from 12.5 to 25 samples per second. This effectively
reduced the computer transport lag. Transport lag was significant contributor to the PIO. Next a nonlinear filter was
applied to the RHC output to attenuate oscillatory inputs and the pitch forward gain was reduced.
Bending filter modifications were applied to all three axes. For the pitch axis only the filter coefficients were
changed however structural changes were made to the roll and yaw axis. In the yaw axis separate fourth order filters
were created for the yaw jets and rudder. This took advantage of a new yaw jet minimum on time thereby reducing
required attenuation resulting in changes to the yaw jet bending filter coefficients. For the roll axis a separate fourth
order bending filter was added to the roll jets. The aileron loop created a second order filter for subsonic flight, a
sixth order filter for 1.0 < Mach < 3.5, and a different sixth order filter for Mach > 3.5. Due to the lower dynamic
pressure a large aileron forward loop gain is required for Mach > 3.5 to achieve the desired transient response. The
10
American Institute of Aeronautics and Astronautics